Effect of the number of positive lymph nodes and lymph node ratio on prognosis of patients after resection of pancreatic adenocarcinoma

Zu-Qiang Liu, Zhi-Wen Xiao, Guo-Pei Luo, Liang Liu, Chen Liu, Jin Xu, Jiang Long, Quan-Xing Ni and Xian-Jun Yu

Shanghai, China

Effect of the number of positive lymph nodes and lymph node ratio on prognosis of patients after resection of pancreatic adenocarcinoma

Zu-Qiang Liu, Zhi-Wen Xiao, Guo-Pei Luo, Liang Liu, Chen Liu, Jin Xu, Jiang Long, Quan-Xing Ni and Xian-Jun Yu

Shanghai, China

BACKGROUND:The prognostic factors related to lymph node involvement [lymph node status, the number of positive lymph nodes, lymph node ratio (LNR)] and the number of nodes evaluated in patients with pancreatic adenocarcinoma after pancreatectomy are poorly defined.

METHODS:A total of 167 patients who had undergone resection of pancreatic adenocarcinoma from February 2010 to August 2011 were included in this study. Histological examination was performed to evaluate the tumor differentiation and lymph node involvement. Univariate and multivariate analyses were made to determine the relationship between the variables related to nodal involvement and the number of nodes and survival.

RESULTS:The median number of total nodes examined was 10 (range 0-44) for the entire cohort. The median number of total nodes examined in node-negative (pN0) patients was similar to that in node-positive (pN1) patients. Patients with pN1 diseases had significantly worse survival than those with pN0 ones (P=0.000). Patients with three or more positive nodes had a poorer prognosis compared with those with the negative nodes (P=0.000). The prognosis of the patients with negative nodes was similar to that of those with one to two positive nodes (P=0.114). The median survival of patients with an LNR≥0.4 was shorter than that of patients with an LNR ＜0.4 in thepN1 cohort (P=0.014). No significance was found between the number of total nodes examined and the prognosis, regardless of the cutoff of 10 or 12 and in the entire cohort or the pN0 and pN1 groups. Based on the multivariate analysis of the entire cohort and the pN1 group, the nodal status, the number of positive nodes and the LNR were all associated with survival.

CONCLUSIONS:In addition to the nodal status, the number of positive nodes and the LNR can serve as comprehensive factors for the evaluation of nodal involvement. This approach may be more effective for predicting the survival of patients with pancreatic adenocarcinoma after pancreatectomy.

(Hepatobiliary Pancreat Dis Int 2014;13:634-641)

lymph node status;

lymph node ratio;

number of positive nodes;

number of total nodes examined;

pancreatic adenocarcinoma;

pancreatectomy

Introduction

Pancreatic cancer has the poorest prognosis of gastrointestinal malignancies, and less than 5% of patients with pancreatic cancer survive more than five years after diagnosis.[1]In the USA, pancreatic cancer is the fourth leading cause of death from cancer.[2]Extended pancreatectomy with dissection of lymph nodes remains the only method for long-term survival and can cure the disease. Several prognostic predictors are available to estimate the survival of patients with pancreatic cancer after pancreatectomy, and these predictors are lymph node metastasis, tumor differentiation and tumor stage.[3]Among them, lymph node metastasis is the most important one to predict the frequency of tumor relapse and metastases.[4,5]

Several studies[6-12]showed that not only the lymphnode involvement (positive or negative) affected the outcomes, but the number of positive nodes and the lymph node ratio (LNR) predicted the survival in patients with gastrointestinal malignancies. Le Voyer et al[9]found that with the same number of nodes involved, the survival was increased when more nodes were analyzed in patients undergoing colectomy for colon adenocarcinoma. To date, the prognostic significance of the number of nodes evaluated, the number of positive nodes and the LNR remain poorly defined in patients with pancreatic cancer, and the data available are limited and controversial.[6,13,14]

The objective of this study was to evaluate systematically the effect of variables related to nodal involvement (nodal status, number of positive nodes and LNR) and the number of nodes evaluated on the survival of patients after surgical resection of pancreatic adenocarcinoma.

Methods

Data were retrieved from the database of our center. All of the patients with pancreatic adenocarcinoma who had undergone surgery between February 2010 and August 2011 were analyzed. The inclusion criteria were resection with radical intent pancreaticoduodenectomy, distal pancreatectomy, total pancreatectomy with a pathological diagnosis of ductal adenocarcinoma and a follow-up for more than 18 months. The exclusion criteria included the perioperative mortality defined as less than one month survival because of complications. Laparoscopic surgery and preoperative chemotherapy or radiotherapy were given. This study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center.

For all the patients, surgical indication was systematically evaluated including a detailed medical history and physical examination, ultrasonography, contrast-enhanced computed tomography, and standard serum laboratory test. Tumor resection was cancelled in cases of the celiac or superior mesenteric artery invasion or distant metastases. The surgical techniques applied by our group during pancreaticoduodenectomy (either the classical Whipple procedure or a pylorus-preserving pancreaticoduodenectomy), distal pancreatectomy and the extent of lymph node dissection were previously described.[6,15,16]Total pancreatectomy was performed when the pancreatic anastomosis could not be safely processed because of the texture of the pancreatic parenchyma or because the residual pancreas was too small to perform anastomosis. Standard lymphadenectomy was performed for pancreatic head cancer along the common hepatic artery, the vena cava, the hepatoduodenal ligament and the right side of the superior mesenteric artery. If the cancer was in the pancreatic body and/or tail, lymph node dissection involved the common hepatic artery (lymph node No. 8), the splenic artery (lymph node No. 11), the inferior edge of the pancreas (lymph node No. 18), and the splenic hilum (lymph node No. 10).[15]Para-aortic lymph node dissections were not routinely performed except the patients who have to achieve intraoperatively a negative retroperitoneal resection margin.

The pathological results were based upon detailed histological examinations of the specimens following the guidelines specified in the dataset for the histopathological reporting. The pathological evaluation of lymph nodes was dependent on lymphadenectomy around the primary tumor. The specimens were fixed in formalin, embedded with paraffin, stained with hematoxylin and eosin and identified as lymph nodes metastases by two pathologists. Tumor stage was evaluated according to the sixth version of the UICC TNM classification.

The data included patient's demographics, operative details, histological characteristics including the origin and classification of the primary tumor, nodal status and patient survival. The data related to the number of total nodes evaluated, the number of positive nodes and the LNR (the ratio of metastatic to examined lymph nodes) were analyzed. The numbers of total nodes examined and metastatic nodes were collected from the pathological reports, and the LNR was obtained according to these data.

The continuous variables, expressed as the median, were analyzed by the Mann-Whitney U test. Spearman's rank-order correlation coefficient test and the Chisquare test were used to investigate the correlation between the clinicopathological parameters and lymph node metastasis. The overall survival probabilities were estimated with the Kaplan-Meier method. A univariate analysis of the differences in the survival curves was made using a log-rank test. A multivariate analysis was performed using the Cox proportional hazards model including the variables that were found to have a P＜0.10 in the log-rank test. Separate multivariable models were constructed for each lymph node variable. The statistical analyses were performed using STATA 12.0. A P＜0.05 was considered statistically significant.

Results

Clinicopathological characteristics

One hundred and sixty-seven patients who met theinclusion criteria were included in the study. There were 75 women and 92 men, with an overall median age of 62 years (range 39-85). The tumors of the patients were mainly located in the head (94 patients, 56.3%). The operations performed included 84 pancreaticoduodenectomies, 69 distal pancreatectomies with splenectomy, and 14 total pancreatectomies. There was no perioperative mortality. Fifty patients (29.9%) had a tumor confined to the pancreas (pT1-2) on the final pathology. The nodal status was positive in 83 patients (49.7%). Nodal positive was significantly related to CA19-9 level, the pT stage and vessel invasion, and not related to the age, gender, tumor location, total bilirubin, nerve invasion, Ki-67 or tumor differentiation, indicating that nodal metastasis occurred with the development of tumor bulk to a certain degree. The clinicopathological characteristics related to the survival analyses are described in detail in Table 1.

Prognostic factors

One hundred and eighteen of the 167 patients died, with a median survival of 11.5 months, and 49 patients were censored. The median survival was 13.6 months in the entire group, and the difference in survival between the node-negative (pN0) and node-positive (pN1) patients was significant (16.8 months for pN0 patients compared with 9.9 months for pN1 patients,P=0.000, Fig. A). The median follow-up time was 12.4 months.

Univariate analysis showed that the following parameters significantly affected the patients' survival:pT stage, TNM stage, vessel invasion, nodal status (positive or negative), the number of positive nodes and LNR. All of the other parameters, such as gender, age, tumor location, tumor diameter, tumor differentiation, preoperative CA19-9 level, preoperative total bilirubin, nerve invasion, Ki-67 (10% as cutoff) and the number of lymph nodes examined had no influence on the survival (Table 1). Multivariate analysis demonstrated that only tumor differentiation, nodal status, the number of positive nodes and LNR were independent prognostic factors. In addition, the c-indices of the four multivariate analysis models were 0.656, 0.681, 0.672, and 0.685, respectively (Table 2, all patients). The variables related to lymph node involvement [nodalstatus, number of positive lymph nodes, LNR (regardless of the cutoff of 0.2 or 0.4)] were correlated with survival on univariate and multivariate analyses. In the pN1 patients, only vessel invasion, the number of positive nodes and LNR ≥0.4 were significantly associated with survival on the univariate analysis (Table 1), and those 3 variables were significant on the multivariate analysis (Table 3).

Table 1.Clinicopathologic characteristics of survivors shown by univariate analysis

Fig.A-E: Prognosis of univariate analysis for the entire cohort stratified by:A: lymph node status (P=0.000);B: the number of positive nodes (P=0.000);C: LNR of 0.2 (P=0.001);D: LNR of 0.4 (P=0.000); andE: the number of nodes examined (P=0.225).F: the differentPvalues of the number of nodes examined by the univariate analysis.

The median number of positive nodes in the pN1 patients was 2 (range 1-27). The median actuarial survival for the patients with positive nodes 0/1-2/≥3 was 16.8, 13.6 and 7.9 months, respectively (P=0.000,Table 1, Fig. B), and the adjusted hazard ratio (HR) of 3 or more nodes was 2.82 (95% CI: 1.74-4.56, P=0.000, Table 2). In the pN1 group, the patients with 3 or more positive nodes had a shorter median survival than those with 1 or 2 positive nodes (P=0.004). In the entire cohort, the patients with 1 or 2 positive nodes had a similar median survival to the node-negative patients (P=0.114). The patients with an LNR=0, 0-0.4, ≥0.4 had median survivals of 16.8, 12.4, and 7.4 months, respectively (P=0.000; Fig. D). When all the pN1 patients were stratified, a cutoff value of 0.4 could predict prognosis independently in the multivariate analysis (P=0.014; HR=1.84, 95% CI: 1.13-3.01; Table 3). However, no significant difference was found in the cutoff value of 0.1 (P=0.147) or 0.2 (P=0.391) by the univariate analysis; this had been frequently reported as significant in many studies.[6,7,17,18]In the entire cohort, a cutoff value of 0.2 was related to survival in the univariate analysis (P=0.001; Fig. C) and multivariate analysis (P=0.013; HR=1.74, 95% CI: 1.13-2.69).

Table 2.Multivariate analysis of the entire cohort

Table 3.Multivariate analysis of the pN1 cohort

The number of median total nodes examined was 10 (range 0-44) for the entire cohort, with the same number of nodes examined for the pN0 and pN1 groups (P=0.829), indicating that there was no marked effect of the number of total nodes examined on the stage evaluation between the patients of the two subgroups. We calculated the P values of the total nodes evaluated, of which the best cutoffs were different in the pN0 and pN1 groups and all patients. We concentrated on the entire group because it was made up of the pN0 and pN1 groups, and the nodal status was unknown during the lymphadenectomy and to the pathologist detecting the nodal specimen. The cutoff values of 2, 8-9, and 11-13 had minimum P values (Fig. F). We created a receiver operating characteristic (ROC) curve, which indicated that the cutoff values of 8.5 and 12.5 had the maximal sensitivity and specificity. It was reported that there was an obvious correlation between the number of examined nodes and the survival in the pN0 patients,[7,19]however, this correlation was not observed in the pN1 group or the entire group. Although it did not reach the significant level, the patients with 12 or more nodes examined tended to have a longer survival (Fig. E).

Discussion

The occurrence of lymph node metastasis is linked to a poor prognosis among patients with pancreatic cancer.[20]The nodal status alone may not have predictive value in the survival of the patients. By univariate analysis, Sohn et al[21]and La Torre et al[22]found that the nodal status was significantly related to patients' survival. However, their multivariate analysis did not confirm this relation. The nodal status was inadequate as an independent factor could be because the prognosis of the node-negative patients is similar to those with 1 or 2 nodes. Riediger and colleagues[23]suspected that the inaccuracy of the nodal status alone in predicting the prognosis might be because of the inadequate lymphadenectomy or inadequate histopathological evaluation.

Emerging evidence suggests that the number of positive nodes and the LNR are more informative than the nodal status for predicting the survival for patients undergoing potentially curative resections of pancreatic cancer.[6,7,14]The number of positive nodes reflects the occurrence and the degree of nodal metastases. Bando et al[10]demonstrated that a higher number of positive nodes was linked to a poorer survival for pN1 patients with gastric cancer. However, this relation is inconclusive currently in patients with pancreatic cancer.[6,7]Slidell et al[7]found that patients with a higher number of positive nodes tended to have a poorer prognosis. They demonstrated that patients with one to three positive nodes achieved a similar survival and that the patients with four to seven positive nodes had a poorer prognosis. Another study[24]demonstrated that the number of positive nodes was an independent parameter for survival, however, no cutoff value was analyzed in their paper. Riediger et al[23]reported that patients with 1 positive node had the same survival as node-negative patients, and the prognosis of patients with negative nodes or 1 positive node was significantly better than that of those with at least 2 nodes involved. These findings were similar to our results. In our study, the HR of the number of positive nodes and c-index of model C was more than nodal status (Table 2, all patients). Studies[6,13]showed that the number of positive nodes was a potential factor for survival after the resection of pancreatic cancer, with more detailed cutoff values to be described and divided in a larger amount of patients.

Several institutions have focused on the LNR to prevent understaging and to explore a desired prognostic evaluation for multiple gastrointestinal malignancies.[25,26]The LNR actually integrates and systematically evaluates the number of positive nodes and the total number of nodes examined. Japanese studies[10,11,27]suggested that LNR was better in predicting the outcome of gastric cancer after surgery compared with the American Joint Committee on Cancer staging, especially for avoiding stage migration. Studies in pancreatic cancer are increasing, indicating that the LNR is more trustworthy than the nodal status alone for predicting survival.[7,14,28]These studies determined that the LNR was significantly associated with the median survival, and the best cutoff values were 0.2 to 0.4.[7,14,28]Our study was consistent with these researches and our cutoff value was 0.2 or 0.4, supporting the previous findings that an increasing LNR is related to a worse prognosis. The adjusted HR of an LNR=0.2 or 0.4 was higher than that of nodal status alone (HR for positive nodal status, 1.61; for LNR≥0.2, 1.74; for LNR ≥0.4, 2.42, Table 2) indicating that the LNR is superior to traditional nodal status alone for predicting survival. The main problem encountered was the various cutoff values involved in the data analysis, which also exists in other gastrointestinal malignancies and hinders the standardization and clinical applica tions.[6,7,17,18,26]On the other side, whether the LNR is superior to the number of positive nodes is still inconclusive. In the study, the HR value of the number of positive nodes was higher than that of the LNR (0.4), which was different from the c-index of the two variables. In addition, we found the colinearity and interaction between the LNR (0.4) and the number of positive nodes. When integrated into one multivariate model (model D, Table 2, all patients), the HR of the two variables was meaningless and significantly different from the two variables in separate multivariable models (model B or model C, Table 2, all patients).

The number of total nodes examined is often regarded as an independent factor for survival in node negative patients.[7,19,29]Previous reports[7,30]have indicated that pN0 patients with more than 12 nodes examined had a better prognosis. Slidell et al[7]identified the differences in the prognosis of 1507 patients with pN0 disease, and confirmed that 12 nodes were the best cutoff value. The difference in survival was up to 7 months (median survival: 1-11 nodes, 16 months vs 12 or more nodes, 23 months).[7]Our study showed that the cutoff values of 2, 8-9, and 11-13 had minimumPvalues for prognosis, and the cutoff values of 8.5 and 12.5 had the maximal values of sensitivity and specificity. Because most institutes recommended the best cutoff between 10 and 15, the National Comprehensive Cancer Network guideline recommends that the minimum number of lymph nodes examined should be between 11 and 17, and we excluded values ＜10 because of our limited data. Our cutoff values were 11-13, and we selected the median value of 12 as the best cutoff. Although we did not achieve significant relationships between the cutoff value and survival time in our study, the median survival time in patients of ≥12 examined lymph nodes tended to be longer than that in patients of＜12 examined lymph nodes.

The median number of total nodes examined in previous studies assessing the prognostic influence of nodal involvement ranged from 7 to 17.[5,8,30]In our study, the median number of evaluated nodes was 10, which may represent a slightly low median number and may be a limitation of this study. It is worth mentioning that the reported median number of total nodes examined in the SEER database was 7,[7]and the SEER database is considered as an authority with a heterogeneous group. The number of median nodes examined was similar between the pN0 and pN1 subgroups in our study, indicating that no stage deviation existed because of the number of total nodes. Our data represented the authenticity of nodal metastasis.

Our study reconfirms that the number of positive nodes and the LNR are better prognostic factors for survival than the nodal status alone after resections for pancreatic cancer. Patients with three or more positive nodes had a poorer survival than those with one or two positive nodes. The prognosis was not different between the patients with negative nodes and those with one to two positive nodes. There were significant correlations between the LNR and the prognostic evaluation with cutoff values of 0.4 for the LNR. This approach may be more effective for predicting the survival of patients with pancreatic cancer after pancreatectomy.

Contributors:LZQ, LGP and YXJ proposed the study. LZQ and XZW performed research and wrote the first draft. LZQ, XZW and LGP collected and analyzed the data. All authors contributed to the design and interpretation of the study and to further drafts. YXJ is the guarantor.

Funding:This study was supported in part by grants from the Sino-German Center (GZ857), Science Foundation of Shanghai (13ZR1407500), Shanghai Rising Star Program (12QH1400600 and 14QA1400900), Fudan University Young Investigator Promoting Program (20520133403) and the National Science Foundation of China (81101807, 81001058, 81372649, 81372653 and 81172276).

Ethical approval:This study was approved by the Ethics Committee of Fudan University Shanghai Cancer Center.

Competing interest:No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

1 Jemal A, Siegel R, Ward E, Hao Y, Xu J, Thun MJ. Cancer statistics, 2009. CA Cancer J Clin 2009;59:225-249.

2 Shi S, Yao W, Xu J, Long J, Liu C, Yu X. Combinational therapy: new hope for pancreatic cancer? Cancer Lett 2012; 317:127-135.

3 Brennan MF, Kattan MW, Klimstra D, Conlon K. Prognostic nomogram for patients undergoing resection for adenocarcinoma of the pancreas. Ann Surg 2004;240:293-298.

4 Luo G, Long J, Cui X, Xiao Z, Liu Z, Shi S, et al. Highlylymphatic metastatic pancreatic cancer cells possess stem cell-like properties. Int J Oncol 2013;42:979-984.

5 House MG, G?nen M, Jarnagin WR, D'Angelica M, DeMatteo RP, Fong Y, et al. Prognostic significance of pathologic nodal status in patients with resected pancreatic cancer. J Gastrointest Surg 2007;11:1549-1555.

6 Massucco P, Ribero D, Sgotto E, Mellano A, Muratore A, Capussotti L. Prognostic significance of lymph node metastases in pancreatic head cancer treated with extended lymphadenectomy: not just a matter of numbers. Ann Surg Oncol 2009;16:3323-3332.

7 Slidell MB, Chang DC, Cameron JL, Wolfgang C, Herman JM, Schulick RD, et al. Impact of total lymph node count and lymph node ratio on staging and survival after pancreatectomy for pancreatic adenocarcinoma: a large, population-based analysis. Ann Surg Oncol 2008;15:165-174.

8 Sierzega M, Popiela T, Kulig J, Nowak K. The ratio of metastatic/resected lymph nodes is an independent prognostic factor in patients with node-positive pancreatic head cancer. Pancreas 2006;33:240-245.

9 Le Voyer TE, Sigurdson ER, Hanlon AL, Mayer RJ, Macdonald JS, Catalano PJ, et al. Colon cancer survival is associated with increasing number of lymph nodes analyzed: a secondary survey of intergroup trial INT-0089. J Clin Oncol 2003;21:2912-2919.

10 Bando E, Yonemura Y, Taniguchi K, Fushida S, Fujimura T, Miwa K. Outcome of ratio of lymph node metastasis in gastric carcinoma. Ann Surg Oncol 2002;9:775-784.

11 Ichikura T, Ogawa T, Chochi K, Kawabata T, Sugasawa H, Mochizuki H. Minimum number of lymph nodes that should be examined for the International Union Against Cancer/ American Joint Committee on Cancer TNM classification of gastric carcinoma. World J Surg 2003;27:330-333.

12 Tepper JE, O'Connell MJ, Niedzwiecki D, Hollis D, Compton C, Benson AB 3rd, et al. Impact of number of nodes retrieved on outcome in patients with rectal cancer. J Clin Oncol 2001;19:157-163.

13 Murakami Y, Uemura K, Sudo T, Hayashidani Y, Hashimoto Y, Nakashima A, et al. Number of metastatic lymph nodes, but not lymph node ratio, is an independent prognostic factor after resection of pancreatic carcinoma. J Am Coll Surg 2010;211:196-204.

14 Pawlik TM, Gleisner AL, Cameron JL, Winter JM, Assumpcao L, Lillemoe KD, et al. Prognostic relevance of lymph node ratio following pancreaticoduodenectomy for pancreatic cancer. Surgery 2007;141:610-618.

15 Sakai M, Nakao A, Kaneko T, Takeda S, Inoue S, Kodera Y, et al. Para-aortic lymph node metastasis in carcinoma of the head of the pancreas. Surgery 2005;137:606-611.

16 Liu C, Long J, Liu L, Xu J, Zhang B, Yu X, et al. Pancreatic stump-closed pancreaticojejunostomy can be performed safely in normal soft pancreas cases. J Surg Res 2012;172: e11-17.

17 Robinson SM, Rahman A, Haugk B, French JJ, Manas DM, Jaques BC, et al. Metastatic lymph node ratio as an important prognostic factor in pancreatic ductal adenocarcinoma. Eur J Surg Oncol 2012;38:333-339.

18 Kawai M, Tani M, Kobayashi Y, Tsuji T, Tabuse K, Horiuchi T, et al. The ratio between metastatic and examined lymph nodes is an independent prognostic factor for patients with resectable middle and distal bile duct carcinoma. Am J Surg 2010;199:447-452.

19 Huebner M, Kendrick M, Reid-Lombardo KM, Que F, Therneau T, Qin R, et al. Number of lymph nodes evaluated: prognostic value in pancreatic adenocarcinoma. J Gastrointest Surg 2012;16:920-926.

20 Long J, Luo GP, Xiao ZW, Liu ZQ, Guo M, Liu L, et al. Cancer statistics: current diagnosis and treatment of pancreatic cancer in Shanghai, China. Cancer Lett 2014;346:273-277.

21 Sohn TA, Yeo CJ, Cameron JL, Koniaris L, Kaushal S, Abrams RA, et al. Resected adenocarcinoma of the pancreas-616 patients: results, outcomes, and prognostic indicators. J Gastrointest Surg 2000;4:567-579.

22 La Torre M, Cavallini M, Ramacciato G, Cosenza G, Rossi Del Monte S, Nigri G, et al. Role of the lymph node ratio in pancreatic ductal adenocarcinoma. Impact on patient stratification and prognosis. J Surg Oncol 2011;104:629-633.

23 Riediger H, Keck T, Wellner U, zur Hausen A, Adam U, Hopt UT, et al. The lymph node ratio is the strongest prognostic factor after resection of pancreatic cancer. J Gastrointest Surg 2009;13:1337-1344.

24 Schwarz RE, Smith DD. Extent of lymph node retrieval and pancreatic cancer survival: information from a large US population database. Ann Surg Oncol 2006;13:1189-1200.

25 Bogoevski D, Onken F, Koenig A, KaifiJT, Schurr P, Sauter G, et al. Is it time for a new TNM classification in esophageal carcinoma? Ann Surg 2008;247:633-641.

26 Rosenberg R, Friederichs J, Schuster T, Gertler R, Maak M, Becker K, et al. Prognosis of patients with colorectal cancer is associated with lymph node ratio: a single-center analysis of 3,026 patients over a 25-year time period. Ann Surg 2008;248: 968-978.

27 Inoue K, Nakane Y, Iiyama H, Sato M, Kanbara T, Nakai K, et al. The superiority of ratio-based lymph node staging in gastric carcinoma. Ann Surg Oncol 2002;9:27-34.

28 La Torre M, Nigri G, Cavallini M, Mercantini P, Ziparo V, Ramacciato G. The glasgow prognostic score as a predictor of survival in patients with potentially resectable pancreatic adenocarcinoma. Ann Surg Oncol 2012;19:2917-2923.

29 Tomlinson JS, Jain S, Bentrem DJ, Sekeris EG, Maggard MA, Hines OJ, et al. Accuracy of staging node-negative pancreas cancer: a potential quality measure. Arch Surg 2007;142:767-774.

30 Berger AC, Watson JC, Ross EA, Hoffman JP. The metastatic/ examined lymph node ratio is an important prognostic factor after pancreaticoduodenectomy for pancreatic adenocarcinoma. Am Surg 2004;70:235-240.

Received April 1, 2013

Accepted after revision January 30, 2014

Author Affiliations: Department of Pancreatic and Hepatobiliary Surgery, Fudan University Shanghai Cancer Center; Department of Oncology, Shanghai Medical College, Fudan University; Pancreatic Cancer Institute, Fudan University, Shanghai 200032, China (Liu ZQ, Xiao ZW, Luo GP, Liu L, Liu C, Xu J, Long J, Ni QX and Yu XJ)

Xian-Jun Yu, MD, PhD, Department of Pancreatic and Hepatobiliary Surgery, Fudan University Shanghai Cancer Center; Department of Oncology, Shanghai Medical College, Fudan University; Pancreatic Cancer Institute, Fudan University, Shanghai 200032, China (Tel: +86-21-64175590ext1307; Fax: +86-21-64031446; Email: yuxianjun88@ hotmail.com)

? 2014, Hepatobiliary Pancreat Dis Int. All rights reserved.

10.1016/S1499-3872(14)60264-2

Published online May 29, 2014.